Two Different Transgenes to Study Gene Silencing and Re-Expression During Zebrafish Caudal Fin and Retinal Regeneration

We used the 500-bp Xenopusef1-α promoter and the 2-kb zebrafish histone 2A.F/Z promoter to generate several independent transgenic zebrafish lines expressing EGFP. While both promoters drive ubiquitous EGFP expression in early zebrafish development, they are systematically silenced in several adult tissues, including the retina and caudal fin. However, EGFP expression is temporarily renewed in the adult during either caudal fin or retinal regeneration. In the Tg(H2A.F/Z:EGFP) line, EGFP is moderately expressed in both the wound epithelium and blastema of the regenerating caudal fin. In the Tg(ef1-α:EGFP) line, EGFP expression is reinitiated and restricted to the blastema of the regenerating caudal fin and colabels with BrdU, PCNA, and msxc-positive cells. Thus, these two ubiquitous promoters drive EGFP transgene expression in different cell populations during caudal fin regeneration. We further analyzed the ability of the ef1-α:EGFP transgene to label nonterminally differentiated cells during adult tissue regeneration. First, we demonstrated that the transgene is highly methylated in adult zebrafish caudal fin tissue, but not during fin regeneration, implicating methylation as a potential means of transgene silencing in this line. Next, we determined that the ef1-α:EGFP transgene is also re-expressed during adult retinal regeneration. Specifically, the ef1-α:EGFP transgene colabels with PCNA in the Müglia, a specialized cell that is the source of neuronal progenitors during zebrafish retinal regeneration. Thus, we concluded that Tg(ef1-α:EGFP)nt line visually marks nonterminally differentiated cells in multiple adult regeneration environments and may prove to be a useful marker in tissue regeneration studies in zebrafish.


INTRODUCTION
Among vertebrates, only certain species, such as specific urodele amphibians and teleost fishes, have the ability to regenerate multiple organs and tissue types. Recently, zebrafish caudal fin regeneration has emerged as an ideal model to further examine vertebrate regeneration due to the simpler anatomical structure of the caudal fin relative to the urodele limb [1]. While publications on teleost fin regeneration α:EGFP transgene is restricted to the mesenchyme, including the forming, proliferating, and distal-most blastema. EGFP expression in the Tg(ef1-α:EGFP) nt line colabels with PCNA, BrdU, and msxc expression in the proliferative blastema. We also found the ef1-α:EGFP transgene was re-expressed in the Müller glial cells during retinal regeneration, which serve as the source of neuronal progenitor cells during photoreceptor regeneration. Thus, the expression of the ef1-α:EGFP transgene in both development and regeneration suggests that it may be a useful tool to visualize nonterminally differentiated cells and may serve as a powerful reagent to screen potential signaling molecules in the regeneration pathways of multiple tissues.

Generation of Tg(ef1-α:EGFP) nt and Tg(H2A.F/Z:EGFP) nt Transgenic Fish Lines
The Tg(ef1-α:EGFP) nt line was generated by coinjecting the pT2KXIG expression plasmid, which contains the 500-bp Xenopus ef1-α promoter upstream of the EGFP reporter gene, with transposase mRNA into 1-4 cell stage zebrafish embryos. Similarly, the Tg(H2A.F/Z:EGFP) nt line was generated by cloning the 2-kb zebrafish H2A.F/Z promoter upstream of EGFP in pT2KXIG and coinjection with transposase mRNA into 1-4 cell stage zebrafish embryos. Six F1 carriers from both lines were identified by EGFP expression, raised and analyzed for EGFP expression at sexual maturity. EGFP-positive adults were out-crossed to the wild-type AB strain. Lines that produced ~50% of the offspring carrying the transgene were raised and analyzed for persistence of the transgene in subsequent generations. Four Tg(ef1-α:EGFP) nt lines consistently produced 50% of offspring carrying the transgene and Southern analysis confirmed that two of these lines contained a single insertion (data not shown). Similarly, two Tg(H2A.F/Z:EGFP) nt lines consistently produced 50% of the offspring carrying the transgene, suggesting that these contain single inserts. All of the single insertion lines have been maintained for six generations.

The ef1-α:EGFP and H2A.F/Z:EGFP Transgene Expression Appears Ubiquitous, but is Silenced in Certain Tissues as Development Progresses
F2 embryos consistently demonstrated maternal expression of the ef1-α:EGFP transgene, as demonstrated by strong EGFP expression prior to the mid-blastula transition (Fig. 1A, n = 4 lines). One line exhibited a mosaic pattern of zygotic expression of the ef1-α:EGFP transgene in the F2 generation (data not shown), and was not analyzed further for this study. In the remaining three lines, strong, uniform zygotic expression of the transgene was observed in all cells by early somitogenesis (~10.5 hpf) (data not shown). This pattern and level of EGFP expression was maintained throughout the body and developing median fin fold until approximately 26 hpf (Fig. 1B). Over the next 96 h, EGFP expression of the transgene was observed in the body, but decreased in the developing fin fold (Fig. 1C). The expression of the ef1-α:EGFP transgene was similarly reduced in the developing retina (Fig. 2). Initially, strong expression was observed in the retinal neuroepithelium at 24 and 42 hpf ( Figs. 2A and B). However, at 60 hpf, expression was dramatically reduced and nonuniform (Fig. 2C). And by 14 days postfertilization (dpf), ef1-α:EGFP transgene expression was below detectable levels in the retina (Fig.  2D), although it persisted in the lens.
Expression of the H2A.F/Z:EGFP transgene was generally less intense than the ef1-α:EGFP transgene (compare Figs. 1 and 3). F2 embryos again demonstrated maternal expression (data not shown, n = 2 lines). H2A.F/Z:EGFP expression was present in the body, retina, and lens through 48 hpf ( Fig.  3A-D). EGFP expression was weakly detected in the developing fin fold at 26 hpf (Fig. 3A), but was absent by 48 hpf (Fig. 3B-D). At 4 dpf, low levels of H2A.F/Z:EGFP expression persisted in the retina (Fig. 3E). However, by 14 dpf, EGFP expression was only detected in the outer and inner plexiform layers (Fig. 3F). There was no significant EGFP expression in the adult retina (Fig. 3G).
To assess which tissues were less prone to silencing of the ef1-α:EGFP and H2A.F/Z:EGFP transgenes, we examined adult transgenic fish for the persistence or absence of EGFP expression (Table  1). We found that both transgenes were silenced in the adult retina (Figs. 2E and 3G), caudal fin ( Fig. 4A and B), and heart atrium (Table 1). Additionally, the ef1-α:EGFP transgene, but not the H2A.F/Z:EGFP transgene, was silenced in the swim bladder and gills (Table 1). EGFP expression from both transgenes persisted in the body muscles, stomach, heart ventricle, and brain (Table 1). EGFP expression was generally stronger in the Tg(ef1-α:EGFP) nt line, especially in the body wall muscles (Table 1).

Tg(ef1-α:EGFP) nt and Tg(H2A.F/Z:EGFP) nt are Differentially Expressed During Caudal Fin Regeneration
Initial analysis of the Tg(ef1-α:EGFP) nt and Tg(H2A.F/Z:EGFP) nt lines revealed that both do not exhibit significant EGFP expression in the adult fins, including the caudal fin ( Fig. 4A and B). Higher-level microscopy revealed very weak EGFP expression in the bony fin rays near the fin girdle, which faded distally ( Fig. 4A and B). This expression was so weak that it could only be observed at high magnification in those fin rays near the girdle that were not obscured by pigmentation.
By 24 h postamputation (hpa), both transgenes re-expressed EGFP in the regenerating fin ( Fig. 4C and D). In the Tg(ef1-a:EGFP) nt line, EGFP expression was first visible in the presumptive-forming blastema and underlying mesenchymal compartment (Fig. 4E). Cryosections taken at 6 and 24 hpa confirm that the ef1-α:EGFP transgene was not expressed in the wound epithelium, but was restricted to the disorganized mesenchyme proximal to the amputation plane and to the overlying and forming blastema ( Fig. 5B and C). During the first 96 hpa, there was a large increase in ef1-α:EGFP expression, during which time EGFP was clearly visualized in the blastema and trailing down to the site of the amputation (Fig. 4F). Following the turnover of the EGFP protein in the newly differentiated cells of the regenerate, by 7 dpa, ef1-α:EGFP expression was restricted solely to the blastema (Fig. 4I). A reduced level of EGFP expression persisted in this pattern through 30 dpa (Fig. 4J), even after fin regeneration appeared to be completed.
In the Tg(H2A.F/Z:EGFP) nt line, EGFP expression was visible by 24 hpa in a diffuse pattern throughout the regenerate ( Fig. 4D and G). This diffuse expression was confirmed in cryosections taken at 6 and 24 hpa. The H2A.F/Z:EGFP transgene, unlike the ef1-α:EGFP transgene, was expressed in both the wound epithelium and forming blastema ( Fig. 5E and F). The H2A.F/Z:EGFP expression increased throughout the first 96 hpa and was visualized throughout the newly-formed tissue (Fig. 4H). By 7 dpa, H2A.F/Z:EGFP expression was restricted to the proximal-most regenerate (Fig. 4K). Unlike the ef1α:EGFP transgene, H2A.F/Z:EGFP expression slowly faded to near background levels by 14 dpa (data not shown) and was not detected at 30 dpa (Fig. 4L).

The ef1-α:EGFP Transgene Coexpressed with PCNA, BrdU, and msxc at 48 hpa
We examined if the ef1-α:EGFP transgene was coexpressed with other known blastema and cell proliferation markers (PCNA, BrdU, and msxc) at 48 hpa in the regenerating caudal fin. The ef1-α:EGFP transgene was expressed in the distal-most and proliferative blastema ( Fig. 6A and E). Similarly, PCNA immunolocalized to the proliferative region of the blastema (Fig. 6B). A merged confocal image of ef1α:EGFP and PCNA immunolocalization revealed that the majority of the ef1-α:EGFP-and PCNApositive cells colocalized ( Fig. 6C and E). A medial compartment of cells located proximal to the mostproliferative blastema, however, contained cells that were solely PCNA positive, solely EGFP positive, and neither ( Fig. 6D and insets). In addition, PCNA-positive cells in the epithelium were not EGFP positive (Fig. 6E, arrowhead), confirming that the ef1-α:EGFP transgene was not expressed in the epithelium. Finally, the distal-most blastema, which has been reported to contain a few cells that do not proliferate at this stage in regeneration [13], contained cells that expressed the ef1-α:EGFP transgene, but not PCNA (Fig. 6E, white arrows).
As an independent method to characterize the expression of the ef1-α:EGFP transgene relative to cell proliferation, BrdU incorporation experiments were performed in the Tg(ef1-a:EGFP) nt line. Caudal fins were partially amputated from adult Tg(ef1-a:EGFP) nt zebrafish and allowed to regenerate for 47.5 hpa, at which point BrdU was injected intraperitoneally. To label only the newly-most formed cells, fins were harvested 30 min later, and wholemount immunolocalization of BrdU was determined. Similar to the results obtained with the PCNA localization, we found many BrdU-positive cells in the proliferative blastema that also expressed the ef1-α:EGFP transgene (Fig. 6F).

The Regenerating Caudal Fin Contains both Methylated and Unmethylated States of the ef1-α:EGFP Transgene
We tested whether de novo DNA methylation was a potential mechanism of the silencing and reexpression of the ef1-α:EGFP transgene in fin regeneration. It was previously shown that EGFP is highly methylated in silenced transgenic mice, and that this methylation can spread to surrounding promoters and transgenes [14,15,16]. In the regenerating zebrafish fin, the ef1-α:EGFP transgene is silenced in the epithelium and only expressed in the underlying blastema tissue. We hypothesized that if methylation was playing a role in the differential expression of the transgene in these two tissues, then the regenerative tissue would contain the transgene in both methylated (silenced) and nonmethylated (not silenced) states. At 4 dpa, the entire regenerative fin tissue (epithelium and blastemas) was harvested and tested for DNA methylation using the standard technique of bisulfite sequencing. Specifically, we analyzed 19 CpG sites surrounding the start codon of the EGFP. We found that the regenerating fin tissue contained two different methylation patterns (Fig. 7), highly methylated (15 or more of the 19 CpG sites methylated) and highly nonmethylated (4 or fewer of the 19 CpG sites methylated). In an independent PCR amplification with different PCR primers, we identified additional highly methylated clones. The methylated CpG sites in the highly nonmethylated clones varied (Fig. 7), making it difficult to determine if any specific sites are potentially critical for transgene expression or silencing. However, the overall level of methylation appeared dramatic.

The ef1-α:EGFP Transgene is Expressed in Dividing Müller Glia, the Source of Neuronal Precursors in the Regenerating Zebrafish Retina
We tested whether the ef1-α:EGFP transgene was re-expressed in other regenerative tissues besides the caudal fin. Zebrafish have the ability to regenerate their retinas following a variety of insults, such as visible light [17,18], heat [19], neurotoxins [20,21,22,23,24], or surgical lesion [25,26]. Although each model results in the loss of a specific subset of neuronal classes, a specialized glial cell, called the Müller glia, serves as a source of neuronal progenitors for regeneration of lost retinal neurons [27,28,29]. Methylation pattern of a region of the ef1-a:EGFP transgene during fin regeneration. A schematic of the ef1-a:EGFP transgene in the Tol2 pT2KXIG element is shown above, with the region of the EGFP reporter gene that was analyzed shown below. DNA samples were obtained from the regenerative tissue of a 4-dpa fin. Two different primer sets were used to amplify an approximately 300-bp region surrounding the EGFP translational start site (marked by an arrow). Each line represents a different PCR subclone; subclones 1-6 were obtained using primer set 1 and clones 7-10 using primer set 2. The state of methylation of each subclone was analyzed using bisulfite conversion of nonmethylated cytosines to uracils and DNA sequencing. Each of the 19 CpG methylation sites in the first primer set region and 17 CpG sites in the second primer set region is represented by a small circle. Notice that both primer sets examine 17 common CpG sites. Empty circles indicate an absence of methylation, while closed circles indicate that the CpG sequence was methylated. Note that two different states of methylation were observed, a high level of methylation in subclones 1 and 2 (top) and a low level of methylation in subclones 3-10 (bottom).
We used the light-lesion model to study the loss and regeneration of photoreceptors in the albino Tg(ef1-a:EGFP) nt line. We previously demonstrated that the Müller glia re-enter the cell cycle within 31 h of light treatment to generate inner nuclear layer (INL) neuronal progenitor [29], which ultimately give , which colabel with the EGFP expressed from the ef1-a:EGFP transgene (panels C and D). The EGFP-expressing cells were determined to be Müller glia based on their morphology, which contain processes that extend from the ganglion cell layer to the outer limiting membrane (panel C), and that the Müller glia are the only PCNA-labeled cells in the INL at this time in regeneration [29]. INL, inner nuclear layer; ONL, outer nuclear layer; GCL, ganglion cell layer. rise to both rod and cone photoreceptors [17]. As shown in Fig. 8, the ef1-α:EGFP transgene was not expressed in the adult retina (Fig. 8A). However, following 36 h of light treatment, EGFP was reexpressed in the Müller glial cells (Fig. 8C). The identity of the EGFP-positive cells as Müller glia was based on their location in the INL, their processes that extend to both the ONL and GCL, and their coexpression of GFAP (data not shown). Further, the EGFP-positive Müller glial cells colabeled with PCNA (Fig. 8D), which is consistent with the ef1-α:EGFP transgene expression in the dedifferentiated neuronal progenitor cells in the regenerating adult retina.

DISCUSSION
Zebrafish is a powerful vertebrate model for the genetic analysis of a variety of developmental processes. In addition, zebrafish have recently become one of the preferred models for the study of tissue regeneration. In particular, there are multiple studies in zebrafish that have focused on regeneration of retinas [17,30], heart tissue [13], and fins [11,31,32,33]. Although researchers utilizing zebrafish have a wider range of genetic tools to study regeneration than many other regeneration model systems, such as the planarian, salamander, and newt, many of the genetic tools routinely used in flies and mice are either currently unavailable or just being developed for use in zebrafish. For example, genetic tools that could be applied to zebrafish regeneration include in vivo electroporation of morpholinos into the regenerating zebrafish fin to knock down the expression of a specific protein [34], the application of Cre-lox for sitespecific recombination [35], and siRNA [36]. The generation and use of cell-specific markers and transgenic lines expressing cell-specific reporters to study regeneration is another area that is needed in zebrafish.
This work describes two lines, Tg(ef1-α:EGFP) nt and Tg(H2A.F/Z:EGFP) nt , that up-regulate EGFP expression in specific cell populations during fin regeneration that could be used -independently or in conjunction with other available genetic tools -to mark particular cell types during the various stages of zebrafish caudal fin regeneration. Following partial amputation of the caudal fin, the first stage in zebrafish fin regeneration is wound epithelium formation. At a regeneration temperature of 33 o C, a thin layer of epithelium forms over the wound by approximately 6 hpa. At this stage in regeneration, expression of the H2A.F/Z:EGFP transgene is expressed in both the wound epithelium and underlying mesenchyme, but not in mature epithelium or mesenchyme (Fig. 5E). During wound epithelium formation, an unknown signal or signals trigger the disorganization of the underlying mesenchyme [37,38], which is followed by osteoblasts and fibroblasts re-entering the cell cycle [8]. Interestingly, during this stage of regeneration, the expression of the ef1-α:EGFP transgene is upregulated in a few mesenchyme cells proximal to the amputation plane (Figs. 4 and 5), which likely represent the osteoblasts and fibroblasts that are re-entering the cell cycle. The dedifferentiated osteoblasts and fibroblasts rapidly divide, and by 24 hpa, a proliferative blastema has formed distal to each bony fin ray. During this stage of regeneration, the H2A.F/Z:EGFP transgene continues to be diffusely expressed in both the wound epithelium and underlying blastema (Figs. 4H and 5F). Expression of the ef1-α:EGFP transgene, in contrast, is restricted to the proliferating blastema and a few mesenchyme cells proximal to amputation plane (Fig. 5C). During fin outgrowth, beginning by 48 hpa, the blastema is subdivided into a distal-most, nonproliferative blastema, composed of only a few cells, and a proximal, proliferative blastema. During this stage, the ef1-α:EGFP transgene is expressed in both the distal-most and proliferating blastema. This was confirmed by ef1-α:EGFP coexpression with PCNA and msxc in the proliferative blastema (Fig. 6).
Taken together, these data suggest that the H2A.F/Z:EGFP transgene marks newly formed wound epithelium and mesenchyme during fin regeneration, while the expression of ef1-α:EGFP transgene may be a useful tool to visualize nonterminally differentiated cells in regenerative tissues. In support of this, we found ef1-α:EGFP expression in the distal-most mesenchyme in SU5402 fins, even though fin outgrowth was completely inhibited (data not shown). In addition, we demonstrated that the ef1-α:EGFP transgene was also re-expressed in the dividing Müller glia cells of the regenerating retina (Fig. 8), which serve as the source of neuronal progenitors for regeneration of lost retinal neurons [27,28,29]. It is important to note that fin regeneration utilizes a blastema-based model of regeneration and regeneration of the light-damaged retina uses Müller glial cells as a source of progenitors. However, in both cases, the ef1-α:EGFP transgene colabeled with PCNA-positive progenitor cells.The mechanism of action behind transgene silencing in these lines is currently unclear. One potential mechanism is de novo DNA methylation of CpG sequences in the transgene. Recently, an elegant study in the mouse model compared the DNA methylation of a ROSA26:EGFP transgene to the endogenous ROSA26 DNA located on chromosome 6 [16]. High levels of methylation were found in the transgene, but low levels were detected in the endogenous promoter region [16]. Because both the EGFP coding sequence and the adjacent ROSA26 promoter were highly methylated, the authors suggested that the EGFP acted as a methylation target that spread to the adjacent ROSA26 promoter [16]. A different group recently reported that the presence of a PGK:EGFP transgene silenced an adjacent LCR:β-globin transgene [14]. Interestingly, when these authors used a modified EGFP that lacked CpG methylation sites, LCR:β-globin expression was partially restored [14]. Similarly, we analyzed regenerating zebrafish fin tissue, which contains wound epithelium, where the transgene is silenced, and regenerative blastemas, where the transgene is expressed. We found both methylated and unmethylated forms of the ef1-α:EGFP transgene in the regenerating fin (Fig. 7). Based on these data and the previous data from the mouse, we hypothesize that DNA methylation and demethylation plays a roll in transgene silencing and re-expression during tissue regeneration.
Although there are many potential uses for these lines, we intend to use them in a subtractive gene microarray experiment designed to reveal cell-and temporal-specific expression of transcripts during fin and retinal regeneration. A previous work describes subtractive hybridization and differential display screens for genes expressed at two time points in zebrafish caudal fin regeneration [6]. Unfortunately, the inability to separate the wound epithelium from the blastema forced the authors to perform in situ hybridizations on a subset of the genes to determine their expression pattern [6]. The transgenic lines described in this work could be used to isolate different cell populations with a fluorescent-activated cell sorter. The corresponding RNAs could then be used in a gene microarray experiment from only a specific subset of cells during the various stages of regeneration. For example, the EGFP-positive cells from the Tg(H2A.F/Z:EGFP) nt line during the latter stages of fin outgrowth would express genes involved in the continuation of fin outgrowth, but not those genes involved in the differentiation of newly formed tissue. Conversely, isolation of EGFP-positive cells from the Tg(ef1-α:EGFP) nt line would express genes involved in blastema formation (at 6-24 hpa) and maintenance (at 48 hpa), while eliminating genes involved in wound epithelium formation and tissue differentiation. Thus, it is feasible that one could design a subtractive microarray experiment from which the gene sets obtained from the entire fin, entire regenerate, H2A.F/Z:EGFP-and ef1-α:EGFP-positive cells (at various stages of regeneration), would reveal a matrix of genes essential to each stage of caudal fin regeneration. Further, by comparing these data to the microarray data sets from regenerating zebrafish retinas [27,29], genes that are essential for general tissue regeneration could be sorted from genes that are essential to regeneration of a specific organ.

Generation of Transgenic Lines
To generate transgenic fish containing the EGFP gene under the Xenopus ef1-α minimal promoter, we used the pT2KXIG plasmid that contains the 500-bp ef1-α promoter upstream of EGFP within a nonautonomous Tol2 transposable element [39]. The plasmid was purified using the Qiagen Maxi Prep Kit (Qaigen; Valencia, CA), followed by phenol/chloroform (1:1) extraction. Purified plasmid (25 ng/μl) was coinjected with in vitro transcribed Tol2 transposase mRNA (25ng/μl) into 1-4 cell stage embryos. The mRNA was produced as previously reported [35]. Embryos expressing EGFP were isolated and raised to adulthood, at which point they were out-crossed to AB (wild-type) adults to determine if any were founder fish that contained the ef1-α:EGFP transgene in their germline. Founder fish were then mated several times to generate families of F1 fish that were used in this study.
To generate transgenic fish containing the EGFP gene under the histone variant promoter (H2A.F/Z), we acquired the H2A.F/Z:EGFP transgene [40]. To clone the H2A.F/Z promoter into the Tol2 EGFP backbone, the following changes were made to the H2A.F/Z-EGFP plasmid. First, the KpnI site in the H2A.F/Z-EGFP plasmid was changed to a XhoI by in vitro mutagenesis using the oligo 5'-CTCGAGGTAC-3'. Second, the NheI site was changed to a SalI restriction site by in vitro mutagenesis using the oligo 5'-CTAGGTCGAC-3'. Using these newly introduced restriction sites, the H2A.F/Z promoter was subcloned into the XhoI-SalI-digested pT2KXIG vector and replaced the ef1-α promoter. It should be noted that this promoter contains a portion of the 5' UTR of the H2A.F/Z gene, which did not prevent EGFP expression. This plasmid was then purified and coinjected with the in vitro transcribed Tol2 transposase mRNA as described above. Founder fish containing the H2A.F/Z:EGFP transgene were isolated and mated as described above.

Fish Husbandry and Collection
All transgenic zebrafish were maintained in the Center of Zebrafish Research at the University of Notre Dame under a 14:10 h light:dark cycle at 28.5 o C [41,42]. Fish were fed three times daily a combination of dry food and brine shrimp. Embryos were collected immediately following fertilization (0-0.5 hpf) and maintained at 28.5 o C. To increase the rate of regeneration approximately twofold, the fish were maintained at 33 o C after caudal fin amputation. Earlier studies demonstrated that the regeneration mechanism was normal at 33 o C, except for the increased rate [5,7].

Wholemount Brightfield and Fluorescent Imaging
Live embryos and adult zebrafish were anesthetized using 2-phenoxyethanol prior to microscopy. Images were taken on a Spot 2 digital camera (Diagnostic Instruments; Sterling Heights, MI) attached to a Leica MZFL III stereomicroscope, with the Leica GFP3 filter for EGFP fluorescent images. All images were edited with Adobe Photoshop 7.0.

Immunolabeling and Confocal Imaging
Following various lengths of time of regeneration, fins were reamputated proximal to the original amputation plane and fixed overnight at 4 o C in a solution of either 4% paraformaldehyde/5% sucrose/1X PBS (for BrdU immunostaining) or 9:1 ethanolic formaldehyde (100% ethanol:37% formaldehyde, for PCNA and EGFP immunostaining). Following fixation, fins were washed (3X, 20 min) in 5% sucrose/1X PBS. Samples were embedded in a heated (55 o C) solution of 5% sucrose/1.5% agar. The tissue was cryoprotected through a series of washes for 10 min each in 30% sucrose, 30% sucrose/TBS, TBS alone. Blocks of fin tissue were embedded in Tissue Freezing Medium (Triangle Biomedical Sciences, Durham, NC) and sectioned at 14 μm. The frozen sections were dried for 2 h at 50 o C and then rehydrated in 1X PBS.
EGFP immunolocalization was performed on fins that were fixed in 9:1 ethanolic formaldehyde for either PCNA coimmunolocalization or following RNA wholemount in situ hybridization (see next section). Proliferating cell localization was assessed using PCNA immunostaining on sectioned fins as described [17]. Briefly, sections were incubated in blocking solution (1X PBS/2% normal goat serum/1% DMSO/0.2% Triton-X 100) for 2 h at room temperature and then overnight at 4 o C in 1:1500 dilution of a rabbit anti-GFP polyclonal antibody (Abcam; Cambridge, MA) and a 1:1000 dilution of mouse monoclonal anti-PCNA antibody (clone PC10, Sigma Chemical; St. Louis, MO) in blocking solution. The sections were washed in 1X PBS/0.05% Tween-20 and then incubated for 1 h at room temperature in 1:500 Alexa Fluor 594 goat antimouse secondary antibody (Molecular Probes; Carlsbad, CA) and 1:500 Alexa Fluor 488 goat antirabbit secondary antibody diluted in 1X PBS/0.05% Tween-20. A nuclear stain, TO-PRO-3 (Molecular Probes), was added to the secondary antibody solution at a 1:750 dilution. Sections were washed in 1X PBS/0.05% Tween-20, 1X PBS and covermounted using ProLong Gold (Molecular Probes). PCNA-positive cells (red nuclei) and EGFP-positive cells (green cells) were analyzed using confocal Z-stack images (depth 10 µm, 1 µm per stack) obtained on a 1024 BioRad confocal microscope.

RNA In Situ Hybridization
Wholemount in situ hybridization was performed as previously described on 24 hpf embryos [43] using a digoxigenin-labeled antisense msxc probe [44]. Following detection of the msxc transcript, fins were postfixed in 9:1 ethanolic formaldehyde and processed for cryosectioning as described above. EGFP immunolocalization was preformed on cryosections as described above and images were obtained on a 1024 BioRad confocal microscope.

Bisulfite Mapping of Methylated and Nonmethylated DNA
Genomic DNA was isolated from the Tg(ef1-α:EGFP) nt line as previously described [35]. An aliquot of 0.5 μg of genomic DNA per sample was subjected to bisulfite conversion using the EZ DNA Methylation-Gold Kit (Zymo Research; Orange, CA). To convert all the nonmethylated cytosines to uracils, 130 μl of CT conversion reagent was added to each DNA sample and incubated in a thermal cycler at 98 o C for 10 min, 64 o C for 2.5 h, and then 4 o C for 20 h. The methylated cytosines fail to react and do not become uracils. The converted DNA was purified using a Zymo-Spin IC column and desulfonated following the protocol in the kit. A yield of approximately 25 ng/μl (250 ng total) was achieved for each sample after column purification. A ~ 300-bp product, which included EGFP's translational start site, was PCR-amplified using Platinum Taq High Fidelity enzyme (Invitrogen) from 125 ng of converted DNA with the following two sets of converted primers: set 1, forward 5' -TTGAGATGAGGATAAAATATTTGAG -3' and reverse 5' -TAATACAAATAAACTTCAAAATCAAC -3'; set 2, forward 5' GGTTGTTGTGTTGTTTTATTATTTTGG -3' and reverse 5' -ATAACTATTATAATTATACTCCAACTTATACC -3'. The cycling conditions were as follows: denaturing at 94 o C for 2 min, followed by 35 cycles of 94 o C for 30 sec, annealing at 52 o C for 1 min, extension at 68 o C for 8 min followed by a single incubation at 68 o C for 10 min at the conclusion of the 35 cycles. PCR fragments were cloned into the TOPO pCR4 vector (Invitrogen) and sent to Sequetech, Inc (Mountain View, CA) for sequencing. The 300-bp pT2KXIG contains either 19 or 17 potential CpG islands (primer sets 1 and 2, respectively), which were scored as nonmethylated if a TG was present and the CpG island (C converted to U) or methylated if the CG was still present at a CpG island. A total of six clones using the first set of primer pair and four clones using the second primer pair were amplified, sequenced, and analyzed.

Constant Light Treatment
Light treatment was performed as previously described [17,18,29]. Briefly, 6-to 9-month-old albino Tg(ef1-a:EGFP) nt zebrafish were dark-adapted for 7 days and then transferred to 2 l clear polycarbonate tanks and placed between four 150-W halogen lamps. Two lamps directly flanked the tank (50 mm from the center of the tank) and two additional lamps were placed at 45 degrees above the tank (70 mm from the center of the tank). This arrangement generated a light intensity of 3,500 lux in the tank. The temperature of the room was tightly regulated such that water temperature did not exceed 33 o C. Fish were light treated for 36 continuous hours, at which point they were euthanized and their eyes harvested and processes for retinal analysis.